Re: When should objects be used?

[Pavel Romashkin <promashkin(at)cmdl.noaa.gov>]

Writing your own objects is another realm I think.  Objects provided by RSI are
useful too.
I am adding my humble opinion to the words of experts above - to say that I have
not had a need yet to write my own objects (and, therefore, don't know how to do
that :-).  The only objects I use are graphics objects.  The reason I preferred to
use object graphics was not the ease of positioning, etc. but  the need to re-use
the data plotted.  Direct graphics is way faster than RSI supplied object graphics
to plot or print.  However, I came finally to the point where I plotted the data
that were a result of quite complex pre-processing, and that data were to be
further modified, but visual inspection was needed.  It was in a widget
application, many intermediate variations were possible.  And the data in direct
graphics is lost, once you plotted it, unless you specifically store it somewhere.
Storing copies of slices from a 3 mb array is possible but I didn't like it.  This
is when I went to objects.  In my primitive application, I use different view
objects to display different modes of operation, and can extract the data from
objects tfor analyses - not to just change the appearance of the plots easily.
This way complicated dataset may be used interactively again and again, without
storing it separately.
I can't pretend to advise people on how to use objects.  This is how object
graphics work for me.
Good luck,
Pavel